Sign in

username:

password:



Not a member?

Search matlab



Search tips

Subscribe to matlab



matlab by Keywords

Atanh | Autocorrelation | Bandpass Filter | C++ | Conv | Database | Deconv | Excel | FFT | Filter | Filtering | FIR | Fourier Transfrom | FSK | Gaussian Noise | Haykin | IFFT | Image | Java | LFSR | LMS | LPC | MEX | OFDM | QPSK | Radix | Random | Sampling | Segmentation | Simulink | Visual Basic | Waveform | Wavelet

Ads

Discussion Groups

Discussion Groups | Matlab DSP | voice recognition system

Technical discussion about Matlab and issues related to Digital Signal Processing.

  

Post a new Thread

voice recognition system - jergang19 - Nov 3 5:09:00 2003



Hi everyone,
I am working on a voice recognition project. A speaker speakes a few
pre-defined
words into the system, and the system matches the audio waveform with
stored
templates. If there is a match, Matlab returns a value to the user.
If there is
no match, user access shall be denied.

Has anyone done this program before? I am very new in Matlab and I
would very
much appreciate it if anyone could mail me their code so that I can
learn from
it.

thanks

jergang19





(You need to be a member of matlab -- send a blank email to matlab-subscribe@yahoogroups.com )

Re: voice recognition system - Ahmad hasan - Nov 10 10:25:00 2003



Heloo Mr. Jergang19

In the Area you chosen, you have to decide taht,

- Your system will be speaker dependent, or independent.
- Your system is for 'Isolated words', 'Connected words',or 'Continous
speach'.

Note:
the simplest one is Isolated words, speaker dependent.

The matching should be not between the stored waveform, incoming one.

Follow the steps.
1- record the voice from MIC. directly with matlab function called
'waverecord.m'
2-deal with the waveforms' .wav' files using ' waveread', and wavewrite'
functions.
3- process the waveform to eleminate the silence from the, and to extract
certain features that are private for any speaker.
4- compare the feauter vectors that are stored with taht coming from speaker
using certain recognition technique.

-To extract features from waverforms use 'Zerocrossing Technique', or
'Energy function'.
-The best recognition techniques are ' MED minimum Eucleadian Distance', '
KNN K Nearest Nieghbor', or ' ANN Artificial Neural Networks'.

Study all topics you dont know listed above, and start programming.

best of wishes
Ahmad hasan




(You need to be a member of matlab -- send a blank email to matlab-subscribe@yahoogroups.com )